Bug Triage Process: Steps, Meeting Agenda and Free Checklist
By GiveFeedback.ioUpdated 9 min read
A bug triage process is how a team turns a pile of new bug reports into a short, ordered list of work. Without one, a client review round produces forty messages in three days, half of them duplicates or change requests, and the developer fixes whatever is at the top of the inbox. The steps, the meeting agenda and the free Word template below give every report a decision, and every accepted bug a severity, a priority and an owner.
What is bug triage?
Bug triage is the process of reviewing new bug reports to confirm them, rate their severity, set their priority and assign an owner, so the team always works on the bugs that matter most. The word comes from emergency medicine, where triage sorts patients by urgency before anyone is treated. In both cases the sorting and the treatment are separate jobs.
Triage doesn't fix anything. Its output is a decision for every report: accepted and scheduled, or closed with a reason. On a website project, that also means separating real bugs from requests for something new, which is where most of the time goes when clients are reviewing.
What are the steps of the bug triage process?
Seven steps, in this order: collect, screen, reproduce, classify, rate severity, set priority, assign and reply. The first three can happen before any meeting; the last four are the decisions the meeting is for.
- Collect every report in one place. Emails, chat messages and calls get turned into reports in the same list, or they never get triaged.
- Screen. Split reports that contain several problems, link duplicates to the original, and send back anything without a page, steps or a screenshot.
- Reproduce. Follow the steps on the reported browser and device. A bug nobody can reproduce gets a specific question to the reporter, not a guess.
- Classify. Is it a bug (it doesn't work as agreed), a change request (the client wants something different) or not a bug at all (it works as designed)?
- Rate severity: what the bug does to one person who runs into it, from critical to trivial.
- Set priority: when it gets fixed, from severity plus the page, the number of people affected and the deadlines.
- Assign and reply. Give it an owner and a target, and tell the reporter what happens next.
Every report leaves triage with one of six decisions:
| Decision | When | What happens next |
|---|---|---|
| Accept | Reproduced and in scope | Severity, priority, owner and target |
| Duplicate | The same problem is already on the list | Link it to the original and close it |
| Needs info | It can't be reproduced from the report | Ask the reporter one specific question; review it at the next triage |
| Not a bug | It works as specified or as designed | Reply with the reason and close it |
| Change request | A new or different feature, not a fault | Move it to the change request list for the client |
| Won't fix | Real, but not worth fixing now | Write down why; close it or keep it as P4 |
Who should attend a bug triage meeting?
Three roles are enough for most website projects: a triage lead who prepares the list and runs the meeting, a developer who can judge severity and the size of a fix, and the person who owns the schedule, usually the project manager or product owner, who sets priority. Everyone else joins only when their bugs are on the list.
| Role | What they bring | When they attend |
|---|---|---|
| Triage lead | Prepares the list, keeps time, records every decision | Always |
| Developer | Confirms severity, estimates the fix, spots duplicates | Always |
| Project manager or product owner | Sets priority against deadlines and the client's goals | Always |
| Tester or QA | Reproduction details and the bugs they found | When their reports are on the list |
| Designer | Whether a visual issue is a bug or the approved design | When there are layout or design reports |
| Client | Business priorities and approval of change requests | For the top items, or through a summary afterwards |
In small teams one person often holds two roles; a freelancer may hold all three. The roles still matter, because each one is a different question: is it real, how bad is it, and when does it get fixed.
How often should you triage bugs?
Match the rhythm to how fast reports arrive. A pattern that works for most website projects:
| Project phase | Triage rhythm | Typical length |
|---|---|---|
| Client review round or pre-launch QA | Daily | 15 to 30 minutes |
| First weeks after launch | Two or three times a week | 15 to 30 minutes |
| Live site in maintenance | Weekly | 15 minutes |
| A critical bug at any time | Immediately, without waiting for the meeting | As long as it takes |
The last row is the exception that makes the rest work. A checkout that fails at 10am doesn't wait for Thursday's triage: whoever finds it escalates it at once, and the next triage only records the decision. Everything else waits for the meeting, which is what protects the developers from constant interruptions.
A 30-minute bug triage meeting agenda
Spend the first five minutes on critical bugs, fifteen on new reports, five on open bugs whose priority may have changed and the last five on open questions and change requests. The agenda only fits in thirty minutes if the triage lead does the screening beforehand.
| Minutes | Item | Output |
|---|---|---|
| 0 to 5 | Critical first: new bugs that look critical, and the status of every open P1 | An owner and a plan for each P1 |
| 5 to 20 | New bugs, one by one: confirm, rate severity, set priority, assign | A decision for every new bug |
| 20 to 25 | Priority changes: open bugs whose deadline, traffic or client plans changed | Updated priorities |
| 25 to 30 | Needs info and change requests: who asks what, what goes to the client | Questions sent, change requests listed |
Before the meeting, the triage lead pulls new reports into one list, merges duplicates, sends incomplete reports back and tries to reproduce anything that looks critical. After it, every bug is updated in the tracker, reporters get their answers and the client gets the change requests as a separate list. The template has both as checklists, plus a decisions log for overrides.
One rule keeps the meeting short: decide, don't solve. As soon as two people start discussing how to fix a bug, note the owner and move on.
Using a bug triage matrix
A bug triage matrix crosses severity with business urgency and suggests a priority for each combination, so the team starts from a default instead of an argument. The template uses this one:
| Severity / Urgency | High | Medium | Low |
|---|---|---|---|
| Critical | P1 | P1 | P2 |
| Major | P1 | P2 | P3 |
| Minor | P2 | P3 | P4 |
| Trivial | P2 | P4 | P4 |
Severity comes from the bug: does it take the site down, break a main feature, look wrong or only bother a perfectionist? Urgency comes from the context: a key page or a fixed date is high, an old page few people visit is low. The definitions of each level, website examples and an Excel version that fills in the suggested priority for you are in our guide to bug severity vs priority.
The matrix suggests; the meeting decides. A typo in the client's company name comes out as P2 and often goes to P1, because the client reviews the site tomorrow and the fix takes two minutes. Write the reason in the decisions log. If the same override keeps appearing, change the matrix.
Bugs, change requests and "works as designed"
On client website projects, the hardest part of triage is deciding what counts as a bug at all. "The logo should be bigger" is a change request. "The menu is in the wrong order" depends on what was approved. "The form doesn't send" is a bug. The reference is always the agreed scope: the approved design, the specification or the acceptance criteria.
- When a report matches the agreed design but the client wants something else, it is a change request. Log it on a separate list, with the page and the screenshot, and send it to the client for approval and, where it applies, a quote. The website change request form has a form and a log for that list.
- When the design never covered the case (a tablet width nobody mocked up, a 40-character product name), treat it as a bug unless fixing it changes the design.
- When the team disagrees, ask the designer before the client. Most "is this a bug?" questions are answered by the approved mockups.
Keeping change requests out of the bug list does two things: it keeps the bug count honest, and it turns unpaid scope into a visible list the client can approve.
Running triage in Jira or another tracker
Give untriaged reports a marker and triage from a saved view. In Jira that can be a needs-triage label or a separate status at the start of the workflow, a filter for open bugs without a priority, and severity labels or a custom field for the result. Our Jira bug report template has a CSV with severity labels, a description structure and two saved filters to start from.
The same idea works in any tool: a Trello list called "To triage", a ClickUp status, a monday.com group. What matters is that new reports land in one visible place and leave it only with a decision. If reports come from your website, they can arrive in that place on their own; the integrations page lists the tools GiveFeedback.io sends them to.
Getting reports that are ready to triage
Screening and reproducing take most of the triage time, and most of that goes into missing information: which page, which browser, what the screen showed. Reports that arrive complete make the meeting short.
With GiveFeedback.io, clients and testers report from the website itself: they click a "Give Feedback" button, mark up a screenshot of the page and describe the problem, with no account. Each report arrives with the page URL, the browser and version, the operating system, the screen resolution and a log of the page's console messages, errors and failed requests, so the developer can often confirm severity from the report alone.
Reports land in a dashboard with Open, Resolved and Archived statuses, and can be created automatically in Jira, Trello, ClickUp, Asana, monday.com, Wrike or Notion, posted to Slack or Discord, or emailed. On paid plans the status syncs both ways with Jira, Trello, ClickUp, Asana, monday.com, Wrike, Slack and Discord, so a bug closed in the tracker is resolved in GiveFeedback.io too. If your developers use Claude Code, the MCP server lets it list a project's reports and read each one in full, screenshot and console log included, which makes a quick first pass for duplicates before the meeting.
Print the agenda, book a recurring 30-minute slot for the next review round and send the first incomplete report back before the meeting starts. Two or three sessions in, the list that reaches the developers will be shorter, ordered and free of surprises.
Frequently asked questions
What is the difference between bug triage and bug prioritization?
Prioritization is one step of triage. Triage starts earlier, by checking that a report is a real, reproducible and unique bug rather than a duplicate or a change request, and it ends later, with an owner, a target and a reply to the reporter. Prioritization is only the part where the team decides when a confirmed bug gets fixed.
What should you do with bugs that can't be reproduced?
Mark them as needing information and ask the reporter one specific question: which page, which browser, what they clicked just before. Keep them on the list for one or two more triage sessions. If nobody can reproduce the bug and it hasn't come back, close it with a note explaining what was tried, so it can be reopened if it happens again.
Should the client attend bug triage meetings?
Usually not the whole meeting. Most of triage is technical sorting the client doesn't need to sit through. Agree the top priorities with them instead: send a short summary after triage with the bugs being fixed first, the questions that need their answer and the change requests that need their approval. Invite them only when a decision depends on their business, such as a launch date.
What is a bug triage matrix?
A grid that crosses bug severity with business urgency to suggest a default priority for each combination, for example a critical bug on a key page giving priority one. It lets the meeting start from a suggestion instead of a debate. The team can override any suggestion, but should write down why, so the matrix improves as the project goes on.
How long should a bug triage meeting take?
Thirty minutes is a good ceiling for most website projects, and fifteen is often enough during a busy review round with daily triage. If it keeps running over, the preparation is missing: duplicates not merged, incomplete reports not sent back, critical bugs not reproduced beforehand. Discussing how to fix a bug belongs after the meeting, between the people doing the fix.
Keep reading
Guides
Website Change Request Form: Free Template, Change Log and Process
A free website change request form (Word) and change log (Excel), how to tell a change from a bug, and a process to quote, approve and ship updates.
- 9 min read
Guides
Jira Bug Report Template: Free CSV Import and Field Guide
A Jira bug report template you can import: a free CSV with four example bugs, a field guide, a description to paste and how to write a bug report in Jira.
- 9 min read
Guides
Website Feedback Form Questions: 45 Examples by Goal (Free Template)
Website feedback form questions for visitors, customers and clients: 45 examples grouped by goal, three ready-made forms and a free Excel question bank.
- 11 min read
Collect visual feedback on your next website
Visual website feedback for agencies, freelancers and their clients.
Free plan - No credit card - Set up in minutes
